7 August 2023 in Transportation Research Part B: Methodological
7 May 2022 in KSCE Journal of Civil Engineering
In a dynamic carpooling system, drivers and riders with their own intended travel plans are matched on short notice. The performance of such a system largely depends on the carpool participants’ travel flexibility (the extent to which a detour is tolerated or the willingness to accept a slightly different drop-off location). To increase travel flexibility, an incentive scheme can be introduced for carpool participants to opt for. For instance, a driver specifies how much she/he expects to be compensated (e.g., $5) if the earliest departure time is shifted to be earlier than the originally scheduled time by a certain amount (e.g., 10 minutes). Similarly, an interested passenger reports the expected incentive to willingly accept a different destination (such as a nearby transit stop or coffee shop) deviating from the request. In this dynamic carpool matching problem with incentives, the following decisions are jointly optimized from the perspective of a carpool matching coordinator: 1) incentive allocations to drivers and riders, 2) assignments of riders to drivers, and 3) vehicle routes of drivers. A case study based on data from Washington, D.C. is conducted to evaluate the potential of the personalized incentives offered to carpool participants in mitigating the environmental impact of transportation (quantified by the reduction of vehicle miles traveled). Two notable findings are reported. First, one dollar of incentive could reduce vehicle miles travelled by 2.88 in one benchmark case. Second, driver incentives are shown to be much more effective than rider incentives under reasonable cost assumptions.
19 June 2025 in Journal of Air Transport Management
25 July 2026 in Accident Analysis & Prevention
1 February 2024 in Intelligent Transportation Infrastructure
In this paper, we aim to address a relevant estimation problem that aviation professionals encounter in their daily operations. Specifically, aircraft load planners require information on the expected number of checked bags for a flight several hours prior to its scheduled departure to properly palletize and load the aircraft. However, the checked baggage prediction problem has not been sufficiently studied in the literature, particularly at the flight level. Existing prediction approaches have not properly accounted for the different impacts of overestimating and underestimating checked baggage volumes on airline operations. Therefore, we propose a custom loss function, in the form of a piecewise quadratic function, which aligns with airline operations practice and utilizes machine learning algorithms to optimize checked baggage predictions incorporating the new loss function. We consider multiple linear regression, LightGBM, and XGBoost, as supervised learning algorithms. We apply our proposed methods to baggage data from a major airline and additional data from various US government agencies. We compare the performance of the three customized supervised learning algorithms. We find that the two gradient boosting methods (i.e., LightGBM and XGBoost) yield higher accuracy than the multiple linear regression; XGBoost outperforms LightGBM while LightGBM requires much less training time than XGBoost. We also investigate the performance of XGBoost on samples from different categories and provide insights for selecting an appropriate prediction algorithm to improve baggage prediction practices. Our modeling framework can be adapted to address other prediction challenges in aviation, such as predicting the number of standby passengers or no-shows.
1 April 2026 in Computational Urban Science